JP5875590B2 - Method and apparatus for printing periodic patterns - Google Patents

Description

  The present invention relates generally to the field of photolithography as used in the creation of microstructures and nanostructures, in particular to the field of photolithography based on the Talbot effect.

  Lithographic creation realizes the deformation of microstructures and nanostructures on the surface. Photolithographic techniques accomplish this deformation by exposing the photosensitive surface to a light field having an intensity distribution corresponding to the desired pattern. The photosensitive surface is usually a thin layer of a sensitive material, such as a photoresist, applied directly to the substrate surface or indirectly over an intermediate layer of another material. It is. The chemical or physical changes that occur in the photosensitive layer as a result of exposure are utilized in subsequent processes to obtain the desired pattern in the substrate material or in an intermediate layer of another material. In the most commonly used photolithography technique, an image of a pattern defined in a mask is projected onto the substrate surface using an optical system.

For many applications, a pattern structure that repeats in one or two dimensions, ie a pattern that includes a unit cell of a periodic pattern, is required. Special photolithography techniques for transferring such patterns from the mask to the substrate are based on the Talbot effect. When the periodic pattern defined on the mask is illuminated with a collimated beam of monochromatic light, the diffraction order in the transmitted light field is reduced to a so-called Talbot plane at a given distance from the mask. Reconstruct the “self-image” of the pattern. For a linear diffraction grating, the self-image separation distance S, known as the Talbot distance, is related to the illumination wavelength λ and the pattern period p by the following equation:
S ≒ 2p ² / λ Formula (1)

  This equation has good accuracy when p >> λ (ie, when the light is diffracted at a relatively small angle), but the approximation gets worse as the absolute value of p gets closer to λ. A mask pattern is printed on the photoresist by placing the photoresist-coated substrate on one of those surfaces (see, for example, “Large area patterning for photonic crystals via coherent diffraction lithography by C. Zanke et al. J. Vac. Sci. Technol. B 22, 3352 (2004)). Furthermore, at an intermediate distance between the self-image planes, a Talbot sub-image having a higher spatial frequency than the pattern in the mask is formed, and this Talbot sub-image is one of those fractional Talbot planes. Printing is performed by placing a substrate coated with a photoresist. The printing results achieved using those techniques are such that the duty cycle of the mask pattern (ie, the size of the structure as a fraction of the structure period) produces a high contrast of the intensity distribution on the Talbot surface or fractional Talbot surface. Improved when selected (see US Pat. No. 4,360,586). It is also known in the prior art to increase the contrast of a Talbot image by creating a periodic pattern in a mask using a phase shift material. Due to the high cost of conventional projection photolithography systems for high resolution periodic patterns, photolithography using Talbot imaging is very important for printing such high resolution periodic patterns. It is advantageous.

  However, the main drawback of Talbot technology is that the intensity distribution of the self-image and sub-image is very sensitive to the distance from the mask, that is, the intensity distribution has a very shallow depth of field. It is. This means that in order to print the diffraction grating correctly, the substrate must be very accurately positioned with respect to the mask. Since the depth of field of the self-image and sub-image depends on the square of the pattern period, this positioning becomes more difficult as the period of the diffraction grating becomes shorter. Furthermore, if the pattern needs to be printed on a very uneven substrate surface, or if there is a topographic structure on the substrate surface, or if the pattern needs to be printed on a thick layer of photoresist, It is considered impossible to obtain the desired result.

Chromatic aberration-free Talbot lithography has recently been introduced as an improved method for printing high-resolution periodic patterns in a cost-effective manner (HH Solak et al., “Achromatic Spatial Frequency Multiplication: A Method for Production of Nanometer-Scale Periodic Structures ", J. Vac. Sci. Technol., 23, pp. 2705-2710 (2005) and US Patent Application Publication No. 2008/0186579). This method offers two significant advantages for lithographic applications. First, the depth-of-field problem encountered when using the classic Talbot method is solved. Second, the spatial frequency is multiplied for a number of pattern types. That is, the resolution of the printed structure is increased with respect to the resolution of the pattern in the mask. In chromatic aberration-free Talbot Lithography (ATL), a mask is irradiated with a collimated beam from a light source having a wide spectral bandwidth, and when a predetermined distance from the mask is exceeded, The transmitted light field forms a so-called still image, whose intensity distribution remains unchanged with further extension of the distance. In the case of a linear grating, the minimum distance d _min from the mask at which it occurs is related to the pattern period p in the mask and the spectral bandwidth Δλ of the irradiation by the following equation:
d _min ≈ 2p ² / Δλ Equation (2)

  Beyond this distance, the Talbot image planes for different wavelengths are continuously dispersed as the distance from the mask generating the still image increases. Thus, by placing a photoresist coated substrate in this region, the substrate is exposed to the entire range of horizontal intensity distributions formed between successive Talbot surfaces for a particular wavelength. That is, the pattern printed on the substrate is the average value or integration (integration) of this range of horizontal intensity distribution, which is substantially insensitive to the vertical displacement of the substrate relative to the mask. This technique therefore achieves a much deeper depth of field than that using standard Talbot imaging, and a much deeper depth of field than using conventional enlargement, proximity printing or contact printing. To realize.

  The intensity distribution in the ATL image derived from a specific mask pattern can be obtained by modeling software that simulates propagation of electromagnetic waves passing through the mask and electromagnetic waves passing through the mask. Such a simulation tool can be used to optimize the design of the pattern in the mask to obtain a specific printed pattern on the substrate surface.

  The ATL method was developed primarily for printing periodic patterns containing unit cells that repetitively appear at a constant period in at least one direction. However, with this technique, the diffraction order forming a particular part of the still image is provided by a portion of the mask whose period is substantially constant, so that the period varies spatially in a sufficiently "slow" manner across the mask. It can be applied to a pattern having Such a pattern can be expressed as quasi-periodic.

  A drawback of ATL is that a light source with a significant spectral bandwidth is required to ensure that the required separation between the mask and the substrate is not inconvenient. Angular divergence of different diffraction orders propagating from the mask results in a spatial offset between the different orders on the substrate surface, which results in incomplete image reconstruction at the pattern edges, which increases the separation distance. It deteriorates more than the accompanying. Fresnel diffraction at the end of the diffraction order also degrades the quality of the end of the printed pattern, and this also degrades as the separation increases. For these reasons, laser light sources having a relatively narrow spectral bandwidth are often not suitable for ATL.

  Since it is difficult to apply non-laser light sources such as arc lamps or light emitting diodes to ATL, it ensures high output of the exposure beam to guarantee high throughput in the production process, high contrast imaging, and Combined with good beam collimation to minimize structural resolution degradation. In order to achieve good collimation from a non-laser light source, spatial filtering of the output beam is required, which generally results in a significant loss of power.

  The benefits of ATL technology are obtained by using different but related technologies disclosed in US Patent Application Publication No. 2008/0186579. In this configuration, a periodic pattern in the mask is illuminated by a collimated beam of monochromatic light, and during exposure the average value of the intensity distribution between the Talbot surfaces is printed on the substrate. Is changed over a range where the distance of the substrate from the mask corresponds to an integral multiple of the separation between successive Talbot image planes. Therefore, the minimum displacement that can be used is equal to the separation distance of the continuous Talbot surface (when integer = 1). With this displacement during exposure, the pattern printed on the substrate is substantially identical to that printed using ATL technology. The above patent document discloses that displacement can be performed continuously or discretely by exposing the substrate at a plurality of discrete positions over the above-mentioned range. When using continuous displacement, the speed of displacement must be constant to obtain the desired average value of the lateral intensity distribution, and when using discrete or stepped displacement The irradiation dose at each discrete location must be equal for similar reasons. This general technique is called Displacement Talbot Lithography (DTL).

  The integrated intensity distribution formed on the substrate using ATL technology and the integrated intensity distribution formed on the substrate using DTL technology are substantially equivalent, both of which are deep depth of field and spatial frequency multiplication for the printed pattern. However, the DTL configuration has the advantage that it can be used with a narrower separation between the substrate and the mask. This reduces the degradation of the pattern edge, and the requirement for collimation is not very strict, so that more effective use of the output from the light source is realized. Furthermore, DTL technology can use a laser light source suitable for the production process. Light from such a light source can be transformed into a well collimated beam with negligible output loss, thereby minimizing structural resolution degradation and maximizing image contrast.

  The structure of a pattern printed using DTL from a specific mask pattern can also be theoretically determined by simulation software.

  Furthermore, the prior art mentions that DTL and ATL can be applied to quasi-periodic patterns, but its details, limitations and drawbacks are not disclosed.

  The disadvantage of DTL technology is that the vertical displacement of the substrate relative to the mask being exposed must correspond exactly to an integer multiple of the Talbot distance. If the displacement is exactly an integer multiple, the integrated intensity distribution that exposes the substrate does not depend on the initial separation between the substrate and the mask, even if the mask and substrate are not exactly flat and parallel. The pattern structure is uniformly exposed. On the other hand, due to, for example, mechanical hysteresis of the displacement actuator or limited stepping resolution, or due to inaccurate synchronization between the exposure time by the illumination system and the displacement of the substrate, If the displacement is not exactly an integer multiple of the Talbot distance, the integrated intensity distribution depends on the initial separation distance. In this case, if the mask and substrate are not exactly flat and parallel, the spatial change in structure size is reflected in the printed pattern, and the mask and substrate are exactly flat and parallel, but their separation distance Are different for different substrates, the printed structure size will vary from substrate to substrate, and in either case is a problem for a particular application. Their sensitivity of the printed pattern to the mask and substrate separation can be reduced by displacing the substrate longitudinally relative to the mask by a large Talbot distance, which reduces structural resolution ( Other problems such as irradiating beam is not collimated well), structural distortion (if displacement direction is not exactly vertical), pattern edge degradation (if gap is too large) This can also be disadvantageous and disadvantageously requires a relatively large range of movement of the mechanical subsystem for displacing the substrate and a relatively fast rate of displacement over a given exposure time.

  A further difficult problem of positioning so that the longitudinal displacement corresponds exactly to an integer multiple of the Talbot distance is generally that the transmitted light field is not exactly periodic in the direction perpendicular to the mask. That is. This will be described below on the basis of two specific examples: a one-dimensional pattern and a two-dimensional pattern. In the case of a one-dimensional periodic pattern, i.e. a linear diffraction grating, the diffraction grating period with respect to the illumination wavelength is such that the zero order diffraction order and the first order diffraction order propagate in the transmitted light field. The resulting interference pattern is exactly periodic in the direction orthogonal to the mask (ignoring the action at the edges of the mask pattern), the self-image plane is well defined, and Exactly separated by Talbot distance. However, if the period of the diffraction grating is such that the second-order diffraction order and possibly higher diffraction orders are also propagated with respect to the wavelength (defined by the zeroth-order and first-order orders). The higher order phase on the self-image plane (as is the case) is not exactly the same as that on the mask plane, and such self-image is not accurately formed and transmitted. The light field is not exactly periodic in the direction perpendicular to the mask. Thus, if the diffraction order is relatively high, the DTL prior art teachings cannot avoid any dependence of the integrated intensity distribution on the initial value of the separation distance between the substrate and the mask, This makes it difficult to print the pattern in a uniform and reproducible manner. In the case of a two-dimensional periodic pattern, there is a more difficult problem with obtaining a precisely periodic light field in the direction orthogonal to the mask. For example, if the period of the pattern components in the orthogonal direction is different, the Talbot distance for each component is also different, so in the general case the transmitted light field will be periodic at any Talbot distance in the propagation direction. No. In yet another example, the pattern structure is arranged in a square grid (i.e., the period of the pattern components in the two directions is equal), and 1 including the diagonally diffracted (± 1, ± 1) orders. If the pattern period is chosen so that only the next diffraction order propagates in the transmitted light field, the different Talbot distances associated with the diagonally diffracted orders will also cause the light in the direction orthogonal to the mask. Degrading the periodicity of the field.

  Yet another difficult problem with prior art teachings of displacement talvolithography is a quasi-periodic pattern having a uniform period but a slowly changing pattern over the pattern area, Or, this displacement Talbot lithography occurs when applied to a mask pattern composed of sub-patterns having a plurality of different sub-pattern periods. With such a pattern, it is not possible to irradiate the complete pattern and to displace the substrate relative to the mask by an exact integer multiple of the Talbot distance that simultaneously satisfies different periods, and is therefore described above. For reasons, it is impossible to print such a pattern uniformly.

  Therefore, the first problem of the present invention is that the periodic pattern of the structure is uniform from the pattern in the mask to the substrate without having to displace the substrate relative to the mask by a distance that exactly corresponds to an integer multiple of the Talbot distance. A method and apparatus for displacement Talbot lithography for reproducible printing.

  The second problem of the present invention is that the relative displacement of the substrate with respect to the mask larger than the large coefficient Talbot distance, the shape of the printed structure, so that the resolution of the printed structure is not unacceptably degraded. And a method and apparatus for displacement Talbot lithography for printing a periodic pattern of structures from a pattern in a mask to a substrate in a uniform and reproducible manner that does not require pattern edge sharpness. is there.

  A third object of the present invention is to construct a structure from a one-dimensional mask pattern having a period with respect to the wavelength of irradiation such that a second or higher order diffraction order occurs in the light field transmitted through the mask. Disclosed is a method and apparatus relating to displacement Talbot lithography for printing a one-dimensional periodic pattern on a substrate in a uniform and reproducible manner.

  A fourth problem of the present invention is that the two-dimensional periodic structure of the structure has a period that is not identical in different directions or from a two-dimensional mask pattern that produces a diagonally diffracted order. A method and apparatus for displacement Talbot lithography for printing a pattern on a substrate in a uniform and reproducible manner.

  A fifth object of the present invention is to make a periodic pattern of structures uniform and reproducible on a substrate from a mask pattern having a period that changes continuously over a mask or a period that changes stepwise. A method and apparatus relating to displacement Talbot lithography for printing on a substrate.

  A sixth problem of the present invention is to make the periodic pattern of the structure uniform and reproducible on the substrate from a mask pattern that does not require precise synchronization between exposure by the illumination system and displacement of the substrate or mask. A method and apparatus relating to displacement Talbot lithography for printing on a substrate.

According to a first aspect of the present invention, there is provided a method for printing at least one of a first periodic pattern of structures and a first quasi-periodic pattern of structures. It has the following steps:
a) preparing a substrate, the substrate having a recording layer provided on a surface thereof;
b) providing a mask comprising at least one of a second periodic pattern of structures and a second quasi-periodic pattern of structures;
c) placing the substrate substantially parallel to the mask and with an initial separation having a first desired value;
d) providing an illumination system for illuminating the mask with substantially monochromatic and substantially collimated light, the intensity of the light field transmitted through the mask exposing the recording layer;
e) illuminating the mask for an exposure time while changing the separation distance by a second desired value and the separation distance change rate, wherein the separation distance change rate and the light Illuminating the mask with an energy density for each incremental change in separation distance that changes at least one of the intensities over a change in separation distance, thereby changing with the change in separation distance;
However, the pattern printed on the recording layer is less sensitive to deviations from the first desired value of the initial separation distance and to deviations from the second desired value of the change in separation distance. Have.

  Most advantageously, the exposure density for each incremental change in separation distance is substantially gradually changed over the displacement distance.

  Advantageously, the exposure density for each incremental change in separation is changed by at least 50%, most advantageously at least 80% over the change in separation.

  Most advantageously, the variation in exposure energy density for each incremental change in separation distance over a change in separation distance has a full width at half maximum (FWHM) that is at least 0.8 times the Talbot distance, If a certain degree of degradation in the uniformity and reproducibility of the printed pattern is acceptable for the particular application involved, use a lower value, reduced to half the Talbot distance, alternatively. Can do. Also most advantageously, the full width of the change should be at least 0.8 times twice the Talbot distance, but less than half of the Talbot distance if the demand for the application is relatively low The lowered value can still alternatively be used.

  Advantageously, the variation in energy density for each incremental change in separation distance over a change in separation distance is substantially within the truncated Gaussian distribution, truncated sinusoidal distribution, and triangular distribution. A distribution that corresponds to one but is at least one of those distributions can alternatively be used.

  Advantageously, the mask-wafer separation is changed at a constant rate continuously over the change in separation, or in the case where the rate of change in separation is changed during the change in separation. It is changed while changing the speed of displacement of at least one of the mask and the wafer. Alternatively, the separation between the mask and the wafer can be changed discretely by changing the separation in a relatively small series of steps over the change in separation, where the separation is continuous. It remains constant over the period of time between steps. For such a step configuration, the energy density per incremental change in separation represents the energy density per step. In cases where the rate of change in the separation is changed with the change in separation, this is by using a constant step distance and by changing the step frequency or using a constant step frequency. And by changing the step distance.

  In the case where the intensity of the illumination is changed during the separation distance, this is done by changing the instantaneous intensity of the beam, or if pulse controlled exposure is used, the frequency of the pulse and This is accomplished by modulating the duty cycle. In the latter case, the intensity of the beam rather represents a time average over at least one pulse period.

  Advantageously, the separation can be varied a plurality of times over the aforementioned changes in separation during the entire exposure, wherein at least one of the rate of change of the separation and the intensity of the irradiation is greater than each of the aforementioned separations. Changed during change.

According to a second aspect of the present invention, there is provided an apparatus for printing at least one of a first periodic pattern of structures and a first quasi-periodic pattern of structures. Includes the following:
a) a substrate having a recording layer provided on the surface;
b) a mask comprising at least one of a second periodic pattern of structures and a second quasi-periodic pattern of structures;
c) means for positioning the substrate substantially parallel to the mask and at an initial separation having a first desired value;
d) an illumination system that illuminates the mask with a substantially monochromatic, substantially collimated light intensity such that a light field transmitted through the mask exposes the recording layer;
e) means for changing the separation distance by a second desired value and at a rate of change of the separation distance;
f) During the illumination of the mask such that the mask is illuminated by an energy density for each incremental change in separation distance that is altered over a change in separation distance, at least one of the rate of change in separation distance and the light intensity. Means to change;
However, the pattern printed on the recording layer is less sensitive to deviations from the first desired value of the initial separation distance and to deviations from the second desired value of the change in separation distance. Have.

  Advantageously, the changing means displaces the actuator at a variable speed or modulates the intensity of the light that illuminates the mask.

  In the latter case, the modification means are preferably provided in the path of the irradiation beam from the light source of the irradiation system which modulates the intensity of the beam irradiating the mask, most preferably modulating under computer control. Includes a variable attenuator. Alternatively, the means for changing the intensity changes the electrical input to the light source such that the output beam output is changed, and most advantageously modulates under computer control.

  Advantageously, the pattern structure in the mask includes a transparent space in a layer of opaque material such as chrome formed on a transparent substrate. Alternatively, the structures are formed on a transparent substrate and are transparent in a layer of transparent or partially transparent material that provides a relative phase shift to the locally transmitted light. Space can be included.

  The one or more patterns in the mask and the printed one or more patterns are one-dimensional patterns (i.e. linear diffraction gratings) or are arranged in a (e.g. square, rectangular or hexagonal grating). Or a combination of a one-dimensional periodic pattern and a two-dimensional periodic pattern. If there are multiple periodic patterns in the mask, the patterns can have the same period or different periods, and their lattice vectors can be in the same direction but different directions. May be.

  Advantageously, the illumination system is fixed with respect to the mask being exposed and forms a uniform illumination beam. Alternatively, the illumination system can scan the beam across the mask so that the time integrated exposure density is uniform across the pattern. In this case, the energy density variation for each incremental change in the separation distance over the distance is repeated sufficiently frequently during exposure so that each point of the mask pattern changes every incremental change in the separation distance over the displacement distance. Subject to the same change in energy density.

  The above aspects and / or the following aspects of the present invention will become apparent and better understood from the several examples described below with reference to the accompanying drawings.

1 schematically illustrates a first embodiment of the present invention for printing a periodic pattern on a wafer in which the rate of longitudinal displacement of the wafer relative to the mask is changed during exposure of the mask. 2 shows a computer simulation of a light field transmitted through a mask used in the first embodiment. Fig. 4 shows the lateral variation of intensity in the time average distribution for printing a wafer using a DTL exposure according to the prior art and the mask of the first embodiment. The sensitivity of the peak intensity in the time average distribution for exposing a wafer using the DTL exposure according to the prior art and the mask of the first embodiment to the average separation distance and displacement distance of the wafer and the mask is shown. FIG. 5 shows the dependence of the time average intensity on the displacement distance at the lateral position on the wafer corresponding to the peak in the integrated distribution using the DTL exposure according to the prior art and the mask of the first embodiment. FIG. FIG. 4 shows the variation in exposure energy density for each displacement increment due to a change in the separation between the wafer and mask used in various embodiments of the present invention. FIG. Average separation distance and displacement of peak intensity in time-averaged distribution for exposing a wafer, using the mask of the first embodiment and the irradiation dose for each displacement increment that varies due to a change in separation distance according to the truncated Gaussian distribution Shows sensitivity to distance. At a lateral position on the wafer corresponding to a peak in the integrated distribution, using the mask of the first embodiment and an irradiation dose for each displacement increment that varies with a change in separation distance according to the truncated Gaussian distribution, The dependence of the time average intensity on the displacement distance is shown. Average separation distance and displacement of peak intensity in a time-averaged distribution for exposing a wafer using the mask of the first embodiment and the irradiation dose for each displacement increment that varies due to a change in separation distance according to the truncated sine distribution Shows sensitivity to distance. At a lateral position on the wafer corresponding to a peak in the integrated distribution using the mask of the first embodiment and the irradiation dose for each displacement increment that varies with a change in separation distance according to the truncated sine distribution; The dependence of the time average intensity on the displacement distance is shown. The average separation distance and the displacement distance of the peak intensity in the time-average distribution for exposing the wafer using the mask of the first embodiment and the irradiation dose for each displacement increment that varies due to the change in the separation distance according to the triangular distribution. The sensitivity to is shown. Time at the lateral position on the wafer corresponding to the peak in the integrated distribution using the mask of the first embodiment and the irradiation dose for each displacement increment that fluctuates due to a change in the separation distance according to the triangular distribution The dependence of the average intensity on the displacement distance is shown. 2 schematically shows a second embodiment of the invention for printing a periodic pattern on a wafer, where the intensity of the beam illuminating the mask is changed during the longitudinal displacement of the wafer relative to the mask. Fig. 6 shows a time average intensity distribution for exposing a wafer using DTL exposure according to the prior art and the mask of the second embodiment. The sensitivity of the peak intensity in the time average distribution for exposing the wafer using the DTL exposure according to the prior art and the mask of the second embodiment to the average separation distance and displacement distance of the wafer and the mask is shown. FIG. 9 shows the dependence of the time average intensity on the displacement distance at the lateral position on the wafer corresponding to the peak in the integrated distribution using DTL exposure according to the prior art and the mask of the second embodiment. The average separation distance and displacement of the peak intensity in the time average distribution for exposing the wafer, using the mask of the second embodiment and the irradiation dose for each displacement increment that varies due to the change in the separation distance according to the truncated Gaussian distribution. Shows sensitivity to distance. At a lateral position on the wafer corresponding to a peak in the integrated distribution, using the mask of the second embodiment and the irradiation dose for each displacement increment that varies due to a change in separation distance according to the truncated Gaussian distribution, The dependence of the time average intensity on the displacement distance is shown. Average separation distance and displacement of peak intensity in time-averaged distribution for exposing a wafer using the mask of the second embodiment and an irradiation dose for each displacement increment that varies due to a change in separation distance according to a truncated sine distribution Shows sensitivity to distance. At a lateral position on the wafer corresponding to a peak in the integrated distribution, using the mask of the second embodiment and an irradiation dose for each displacement increment that varies with a change in separation distance according to the truncated sine distribution; The dependence of the time average intensity on the displacement distance is shown. The average separation distance and displacement distance of the peak intensity in the time-average distribution for exposing the wafer, using the mask of the second embodiment and the irradiation dose for each displacement increment that varies due to the change in the separation distance according to the triangular distribution. The sensitivity to is shown. Time at the lateral position on the wafer corresponding to the peak in the integrated distribution using the mask of the second embodiment and the irradiation dose for each displacement increment that varies due to the change in the separation distance according to the triangular distribution The dependence of the average intensity on the displacement distance is shown.

  In the following, reference is made to FIG. 1, in which a first embodiment of the invention is shown. Argon ion laser 1 has a wavelength of 363.8 nm and a diameter of ˜2 mm and is a substantially monochromatic beam of light in a single transverse mode (ie with a Gaussian intensity profile) and in multiple longitudinal modes. 2 is emitted. The light is plane polarized and the polarization vector is orthogonal to the plane of the figure. Such lasers are commercially available, for example, by Newport Corporation (particularly the product of BeamLok lasers) or Coherent Inc. (particularly the product of Innova Saber lasers). After passing through the electronically operated shutter 3, the diameter of the beam 2 is expanded by a beam expander 4 equipped with a set of lenses, and thus the intensity of the resulting beam using the beam converter 6 The profile can easily be changed from a Gaussian distribution with a substantially uniform distribution over the central part of the beam. Suitable beam transducers are commercially available, for example by Moltech GmbH (particularly a product of the piShaper product). The output beam of the beam converter 6 passes through the second beam expander 8. This second beam expander 8 forms a collimated beam, ie a collimated beam, whose central uniform part has a diameter larger than the size of the pattern to be printed. The beam is reflected by the mirror 10 towards the mask 12 so that the parallel beam 11 is incident at right angles and the pattern 13 on the mask 12 is illuminated substantially uniformly. On the lower surface of the mask, there is provided a one-dimensional periodic pattern 13 (that is, a linear diffraction grating) having a period of 800 nm composed of a plurality of opaque lines and a plurality of transparent spaces. Pattern 13 is created in a layer of chromium on a thick (eg, 0.25 ") fused silica substrate using standard electron beam mask making techniques. Although only lines and spaces are shown, it is understood that there are many orders of magnitude in number and that the mask pattern 13 typically has dimensions of centimeters. It is firmly fixed to a frame (not shown).

  Below the mask 12 is a wafer 14 on which an upper surface of the wafer 14 is spin-coated with a ˜1 μm thick layer of standard i-ray sensitive photoresist 15. The wafer 14 is mounted on the vacuum chuck 16. The vacuum chuck 16 is attached to a mechanical positioning system 17 that is configured to position the wafer 14 substantially parallel to and in the vicinity of the pattern 13 in the mask 12. Works with multiple actuators. Those actuators are displaced by the control system 18. The actuators also preferably have three piezoelectric transducers (PZT), each of which is a closed loop control of their respective displacements to minimize displacement errors caused by hysteresis and drift. And has an integrated strain gauge or capacitive sensor, and advantageously has a long travel distance, for example 100 μm. For example, the wafer 14 is adjusted parallel to the mask 12 and in the vicinity of the mask 12 using known and equal thickness reference spacers provided on different surfaces of the wafer 14. The separation distance between the wafer 14 and the mask 12 is generally initially set to a value of 20 μm. For displacement Talbotlithography, this parameter should be small enough but not critical so that the angular range in the illumination beam due to non-perfect local collimation cannot tolerate the resolution of the printed pattern It does not deteriorate as much. Most advantageously, the vertical displacement of the wafer 14 required during the exposure operation is reduced to the surface of the wafer 14 until any lateral displacement component during exposure is reduced with respect to the period of the printed pattern. In order to be accurately perpendicular, the positioning system 17 should cooperate with a guide device or equivalent mechanism (as is well known to those skilled in the art of standard precision positioning systems). In addition, the control system 18 allows the wafer 14 to be displaced longitudinally with a speed that is changed during displacement according to a predetermined profile as required during the exposure operation. The operation for adjusting the parallelism and separation between the wafer 14 and the mask 12 may be performed by manual or automatic means (eg, stepping) to produce a coarse resolution longitudinal displacement of the wafer 14 relative to the mask 12. It can be simplified by additionally integrating the motor) into the mechanical positioning system 17.

Irradiation of the diffraction grating pattern 13 with a period of 800 nm using the collimated beam 11 having a wavelength of 363.8 nm generates zero-order and first-order diffraction orders, and this diffraction order is a series in the transmitted light field. Interferes with the formation of self-image surfaces. The separation distance S ₀₁ is
S ₀₁ = λ / (cos θ ₀ −cos θ ₁ ) (3)
Obtained by. However, θ ₀ and θ ₁ are the diffraction angles of the 0th order and the 1st order, respectively.

From equation (3), S ₀₁ ≈3.3 μm. However, due to the large diffraction angles of the first and second order beams (˜27 ° and ˜66 °, respectively), the second order diffraction orders are more accurate than the zeroth and first order beams in their planes. Therefore, the self-image of the mask pattern 13 is not accurately formed, and moreover, it is not accurately positioned on a clear Talbot surface. As a result, it is very difficult to uniformly print the pattern on the wafer 14 when the mask pattern 13 is exposed at this irradiation wavelength using displacement Talbot lithography according to the prior art. The sensitivity of the line width of the printed pattern to changes in the separation distance between the mask 12 and the wafer 14 can be evaluated by computer simulation of the exposure process. Such computer simulations advantageously use standard theoretical techniques for calculating the propagation of electromagnetic waves through periodic microstructures and uniform media, such as the finite difference time domain (FTDT) method. domain) or rigorous coupled wave analysis (RCWA). Commercially available software or free software such as GSolver (Grating Solver Development Co. for RCWA) or MEEP (Massachusetts Institute of Technology for FTDT) can be used. A diffracted light field transmitted through the mask pattern 13 used in this embodiment, which is irradiated by normal incidence of a beam 11 having a wavelength of 363.8 nm, is simulated, and the result is shown in FIG. Has been. FIG. 2 shows a part of the light field extending from the mask 12 to a distance of 8 μm over the width of one period of the diffraction grating pattern 13. Although the relative phase of the secondary beam on the self-image plane changes, the intensity distribution is not exactly periodic as the distance increases, but the ~ 13 μm self of the pattern 13 in the mask 12 It can be seen from the figure that regular intervals of images are formed. The effect of applying the DTL method to the illustrated diffraction grating pattern 13 can be determined by integrating or integrating the light field over a range of distances corresponding to integer multiples of the separation distance of successive self-image planes. it can. Using a wafer 14 at an initial distance of 20 μm from the mask 12 and a light field integrated over a double Talbot distance range (ie, over 6.6 μm), the time average intensity distribution exposing the photoresist 15 is bright. A periodic pattern of lines, the period of which is half the period of the pattern in the mask, as is generally obtained when DTL is applied to a one-dimensional mask pattern. The change in intensity across this distribution (in a direction orthogonal to the line) is shown in FIG. FIG. 3 shows the change over the distance corresponding to one period of the mask pattern 13. Even if the mask 12 and the wafer 14 are not exactly flat or parallel, in order to print the pattern from the mask 12 to the wafer 14 with uniform reproducibility, the peak intensity in the integrated distribution is It must be not sensitive to the local initial separation of the mask 12 or sensitive to deviations from the desired value of the actual displacement of the wafer. The dependence of the peak intensity on the initial separation distance and the dependence of the peak intensity on the displacement distance can be evaluated by computer simulation. The result of such an evaluation for the irradiation pattern considered here is shown in FIG. FIG. 4 shows the absolute value of the intensity peak plotted in relation to the average separation distance between the wafer 14 and the mask 12 during exposure for various displacement distances of the wafer 14. As can be seen from FIG. 4, when the displacement distance is 6.6 μm (that is, twice the Talbot distance), the peak intensity varies greatly to ˜7% with the variation of the average separation distance, whereas 6.7 μm. And at a displacement distance of 6.8 μm, the intensity variation is 4.5% and 4%, respectively. Further determining the sensitivity of the printed pattern to the absolute value of the displacement by determining the dependence of the accumulated intensity or time average intensity on the displacement at the point having the lateral position corresponding to the intensity peak in the accumulated distribution. Can be evaluated. In FIG. 5, the integrated intensity reaches a maximum value when the displacement is ˜4 μm, continues to fluctuate significantly as the displacement distance further increases (up to ± 13%), and finally has a displacement of> 30 μm. Results are shown to confirm that a reasonably stable value has been reached. These results show the difficulty and limitations of applying DTL technology to the patterns considered here where high line width uniformity and high repeatability are required across the printed pattern and from wafer to wafer. It has been demonstrated.

In this embodiment of the invention, the wafer 14 is displaced relative to the mask 12 at a constant speed during exposure so that the range of intensity distribution between the Talbot planes is evenly recorded in the photoresist layer 15. instead of, instead, the irradiation dose for each displacement increment (displacement increment where small relative to the overall displacement) as E _G is changed with distance d of the wafer 14 and mask 12 in accordance with truncated Gaussian distribution The speed of displacement is modulated during displacement:
E _G (d) = E ₀ exp {− (d−d ₀ ) ² / 2σ ² } where | d−d ₀ | <tσ Equation (4)
Where E ₀ is a constant, exp {} is an exponential function, d ₀ is the average separation during displacement, σ is the standard deviation of the Gaussian distribution, and t is the truncation parameter of the Gaussian function, | X | represents the absolute value of x.

Using the apparatus of FIG. 1, this profile of incremental dose that varies with separation distance changes the wafer 14 at a speed v _G that the actuator varies inversely with respect to the energy density change represented by equation (4) during exposure. Can be accomplished by programming the control system 18 to move toward or away from the mask 12:
v _G (d) = kexp {(d−d ₀ ) ² / 2σ ² } where | d−d ₀ | ≦ tσ Equation (5)
Here, k is a constant.

  A function representing the required position of each actuator depending on time can be derived directly from equation (5).

To obtain the desired exposure result, it is recommended that t≈2, in which case E _G (d) drops to ˜13% of its maximum value at the end of the separation range, but the application requirements Other values can be used depending on. Furthermore, although it is recommended that σ be selected to substantially correspond to half of the Talbot surface separation, this value can be adjusted according to the specific requirements of the application. For t = 2 and σ = T / 2 (T is the Talbot plane separation), the resulting curve representing E _G (d) is shown in FIG. 6 (indicated by “Gauss”). Have been). In order to ensure that the photoresist 15 is exposed only to the light field transmitted through the mask 12 during the exposure defined by the above equation (4), the shutter 3 is advantageously opened and closed by the control system 18. The start and end of exposure are synchronized with the movement of the actuator. The constant k is simply a magnification, and the total irradiation dose that irradiates the layer of photoresist 15 should be selected in combination with the intensity of the irradiation beam in order to form the desired structure in the photoresist 15 according to its development. It is. The irradiation dose is advantageously optimized experimentally by printing a plurality of wafers with different doses and evaluating the printing results.

  The effect of the displacement of the wafer 14 during exposure at a variable speed represented by the equation (5) on the time-average intensity distribution for exposing the photoresist 15, in particular, the distribution with respect to the average separation distance between the wafer 14 and the mask 12. The effect on peak intensity dependence and peak intensity dependence on the deviation of the actual displacement from the desired value can be evaluated by computer simulation. σ = 1.65 μm, 1.675 μm and 1.7 μm, and t = 2, so that the displacement distance of the wafer is 6.6 μm, 6.7 μm and 6.8 μm, respectively, 15 μm to 25 μm The result of such a simulation for the illumination pattern with respect to the average separation is shown in FIG. Thereby, the fluctuation of the peak intensity due to the change of the average separation distance can be estimated to be ˜2.5%, ˜2% and ˜1.5%, respectively, with respect to the three values of displacement. Is much lower than the corresponding value previously determined for DTL exposure by and has been demonstrated experimentally. The sensitivity of the printed pattern with respect to the absolute value of the displacement can be further evaluated by determining the dependence of the integrated intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the integrated distribution. FIG. 8 shows that the average intensity during exposure reaches a maximum at a distance corresponding to about twice the Talbot distance (ie 6.6 μm) and oscillates with an amplitude of <± 1.5% as the distance further increases. The results can be seen. These residual variations can be further reduced by adjusting the parameters of the truncated Gaussian distribution, if necessary. The results show that the exposure of the mask with exposure density per displacement increment that fluctuates with the variation of the separation distance according to the truncated Gaussian distribution provides greater uniformity and reproduction than that obtained using prior art DTL. It has been proved that it is possible to realize a printed pattern with the characteristics.

Uniformity and reproducibility of the significant improvement of the printed pattern, albeit at the same device in this embodiment, irradiation dose E _S for each displacement increment, in response to the truncated sine function, the wafer 14 and mask This is accomplished by an apparatus having a control system 18 that is programmed to modulate the displacement rate of the actuator during exposure to vary with a change in the twelve separation distance d:
E _S (d) = E ₀ COS ² {Π (d−d ₀ ) / 2L} where | d−d ₀ | ≦ tL Equation (6)
Where d ₀ is the average separation during exposure, 2L is the period of the sinusoidal variation, and t defines the truncation of the sine distribution, thus determining the full displacement of the wafer 14 during exposure. .

This profile of incremental dose variation due to separation distance variation can be obtained by modulating the wafer displacement speed v _S according to the following equation:
v _S (d) = ksec ² {Π (d−d ₀ ) / 2L} where | d−d ₀ | ≦ tL Equation (7)
Here, k is a constant.

In order to limit the maximum speed of the required displacement of the actuator, with respect to the irradiation pattern, L should be chosen to correspond to the Talbot distance T and t is near 1 but less than 1, for example It is recommended that it be 0.9. The shape of the curve representing E _S (d) for L = T and t = 1 is shown in FIG. 6 (indicated by “sine”). When these values are used, the variation that occurs as a result of the irradiation dose for each displacement increment due to the expansion (or reduction) of the separation distance of the wafer 14 approximates the above-described truncated Gaussian distribution. The control system 18 should advantageously automatically open and close the shutter 3 during the displacement of the actuator so that the photoresist is exposed only to the light field transmitted through the mask for the required distance of displacement. . The magnification k should be selected in combination with the intensity of the irradiation beam so that the total irradiation dose for irradiating the photoresist 15 forms a desired structure in the photoresist 15 according to its development. The irradiation dose is advantageously optimized experimentally by printing multiple wafers using different doses and evaluating the printed results.

  The computer simulation shows the sensitivity of the peak intensity in the integrated distribution or time average distribution for exposing the photoresist 15 to the variation in the average separation distance between the wafer 14 and the mask 12, and the sensitivity to the deviation of the displacement distance of the wafer 14 from the optimum value. The same is done to find When L = 3.3 μm, 3.35 μm and 3.4 μm and t = 1, the wafer displacement distances are 6.6 μm, 6.7 μm and 6.8 μm, respectively, and the results for the associated mask pattern are respectively It is shown in FIG. This allows the peak intensity variation to be estimated to be ~ 0.03%, ~ 0.6%, and ~ 1% for the three values of displacement, respectively, i.e. preceding the DTL exposure according to the prior art. Is much lower than the corresponding value determined. The sensitivity of the printed pattern with respect to the absolute value of the displacement can be further evaluated by determining the dependence of the integrated intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the integrated distribution. In FIG. 10, it can be seen that the integrated intensity reaches a maximum when the displacement is approximately twice the Talbot distance (ie, 6.6 μm), and fluctuates with an amplitude of <± 1% as the distance further increases. Results are shown. The results show that the exposure of the mask with exposure density for each displacement increment that varies with the change in separation according to the sinusoidal distribution provides much higher uniformity and higher than that obtained using prior art DTL. It has been proved that a one-dimensional pattern to be printed is realized with reproducibility.

Similar improvements in uniformity and reproducibility of the printed pattern, the irradiation dose E _T for each displacement increment, in response to the truncated triangular function, displacement of the actuator during exposure to vary with changes in the distance d Realized by modulating the speed of:
E _T (d) = E ₀ (L− (| d−d ₀ |) where | d−d ₀ | ≦ tL (8)
Where d ₀ is the average separation, 2L is the width of the trigonometric function (not truncated), and t defines the truncation of the triangular function, thus determining the displacement distance of the wafer 14 during exposure. .

This profile of incremental dose variation with variation in separation can be obtained by modulating the wafer displacement velocity v _T according to the following equation:
v _T (d) = k / (L | d−d ₀ |) where | d−d ₀ | ≦ tL (9)
Here, k is a constant.

In order to limit the maximum speed of the required displacement of the actuator, with respect to the irradiation pattern, L should be chosen to correspond to the Talbot distance T and t is near 1 but less than 1, for example It is recommended that it be 0.9. The shape of the curve representing E _S (d) for L = T and t = 1 is shown in FIG. 6 (indicated by “triangle”). When these values are used, the variation that occurs as a result of the exposure density per displacement increment due to the increase (or reduction) of the separation distance still approximates the truncated Gaussian distribution described above. The control system 18 should advantageously automatically open and close the shutter 3 during the displacement of the actuator so that the photoresist 15 is exposed only to the light field transmitted through the mask 12 during the required range of displacement. It is. The magnification k should be selected in combination with the intensity of the irradiation beam so that the total irradiation dose for irradiating the photoresist 15 forms a desired structure in the photoresist 15 according to its development. The irradiation dose is advantageously optimized experimentally by printing multiple wafers using different doses and evaluating the printed results.

  Computer simulation exposes the photoresist 15 and the sensitivity of the peak intensity in the resulting integrated or time-averaged distribution to variations in the average separation between the wafer 14 and the mask 12 as well as the wafer 14 from a desired value. A similar procedure is performed to determine the sensitivity to the deviation of the displacement distance. For L = 3.3 μm, 3.35 μm, and 3.4 μm, and t = 1, the wafer displacement is 6.6 μm, 6.7 μm, and 6.8 μm, respectively, and the results produced for the associated pattern are shown in FIG. 11. Thereby, the variation in peak intensity due to the change in average separation can be estimated to be ˜1% for each of the three values of displacement, i.e. the corresponding previously obtained for DTL exposure according to the prior art. Much lower than the value. The sensitivity of the printed pattern to the absolute value of the displacement is further evaluated by determining the dependence of the accumulated intensity or time average intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the accumulated distribution. be able to. FIG. 12 shows that the average intensity reaches the maximum when the displacement is about twice the Talbot distance (ie, 6.6 μm), and oscillates with an amplitude of <± 1% as the distance further increases. The results that can be seen are shown. The results show that the exposure of the mask with exposure density at varying displacement increments that vary with the change in separation according to the truncated triangular distribution is much higher than that achieved using prior art DTL. It has been proved that a one-dimensional pattern to be printed can be realized with good performance and reproducibility.

  It can be seen from FIG. 6 that truncation Gaussian variation, truncation sinusoidal variation and truncation triangular variation of the incremental energy density are similar due to the change in separation distance described in this embodiment. That is, in each case, the full width of the profile is approximately twice the Talbot distance, and the full width at half-maximum (FWHM) is approximately the Talbot distance. In all cases, the incremental energy density changes substantially gradually over the displacement distance, and the minimum value of the initial and end values of the separation distance is> 80% of the maximum value. Thus, similar improvements in the uniformity and reproducibility of the printed pattern can be achieved using properties similar to those shown, for example, an incremental dose variation profile with an appropriate trapezoidal distribution. Can be achieved. Furthermore, logical and experimental results show that the higher uniformity and repeatability of the printed pattern is from the values suggested in FIG. 6 (with respect to the Talbot distance for the associated periodic pattern). Obtained by incrementing full width and FWHM. However, the greater displacement of the wafer 14 with respect to the mask 12 during exposure requires higher accuracy in the orthogonality of displacement with respect to that plane (in order not to reduce printing resolution) and has a larger range of movement (one (Or more) actuators are required and not necessarily desired.

  In general, to obtain very good uniformity and reproducibility of the printed results, the FWHM of the curve representing the change in exposure energy density for each displacement increment of the separation distance over the displacement distance is at least 0.8 of the Talbot distance. A value that is doubled but smaller, down to half the Talbot distance, can be used alternatively with lower uniformity and reproducibility results, and the distribution The total width should advantageously be at least 0.8 times twice the Talbot distance, but also smaller values that are reduced to the Talbot distance are used to accommodate relatively demanding applications You can also.

  With respect to the separation at which the maximum energy density per displacement increment occurs (as is the case for all curves shown in FIG. 6), the curve representing the variation in energy density from increment to increment due to a change in separation is symmetrical. Or at least substantially symmetric is advantageous, but not necessary. Thus, advantageously, the separation at which the maximum energy density per displacement increment occurs is at a midpoint between the initial and final values of the separation, or at least substantially at the midpoint.

  Most advantageously, the initial value and final value of the separation distance between the mask 12 and the wafer 14 is at least 80% of the energy density value per displacement increment over the range of separation distances. Few. However, a similar variation in energy density per displacement increment over the displacement distance can also be used, but still much better than the uniformity and reproducibility of patterns printed using prior art DTL exposure. Despite this, the uniformity and reproducibility of the printed result is low.

  In the above-described embodiment, the variation of the displacement speed of the wafer 14 relative to the mask 12 is achieved by changing the displacement speed of the wafer 14, but in another embodiment of the present invention, the same Effects and results can be achieved by varying the speed of displacement of the mask 12 as well. In this case, a suitable mechanical system that cooperates with one or more actuators and associated control system should displace the mask 12 longitudinally at a variable speed during exposure.

  By exposing the mask with an energy intensity for each displacement increment of a separation distance that varies over the displacement distance (eg, opening the shutter) so that the incremental exposure density at the edge of the region is small with respect to the incremental exposure density at the center of the region. An additional advantage is provided that the synchronization accuracy required between the start and end of the exposure (defined by the closure) and between the start and end of the displacement can be reduced. In contrast, using DTL according to the prior art, the energy density per displacement increment at the two ends of the scan range is the same as the energy density per displacement increment at the center, so that the printed pattern is exposed. It becomes even more sensitive to synchronization errors between the mechanism and the displacement mechanism.

  In the embodiment described above, the wafer displacement with respect to the mask during exposure is performed in a single direction with the required variation of speed over the required range, but in another embodiment of the invention, Alternatively, multiple displacements of the wafer over a range can be performed during exposure, in which case each displacement is in accordance with the above teachings and advantageously the direction of displacement is between successive displacements. Inverted. Repeating the same movement of displacement during exposure has proven to be the same pattern printed on the wafer. Using repeated displacement over the scan range during exposure further reduces the sensitivity of the printed pattern to synchronization errors between the total exposure time defined by the illumination system and the total time the wafer is displaced. And therefore advantageous.

  In another embodiment of the invention using the same apparatus as used in the first embodiment, or using an equivalent apparatus, the wafer is only the same final distance during exposure. Although displaced with respect to the mask, it is displaced using a stepped motion in which the wafer is displaced in successive steps and with varying delay times between steps. Expose the wafer by selecting a small step size with respect to the final distance of displacement and by selecting a delay time to be changed in response to the required variation in incremental dose due to changes in separation distance It is understood that the integrated intensity distribution approximates the integrated intensity distribution generated by the first embodiment, and the printing results are substantially the same.

  Reference is now made to FIG. FIG. 13 shows a second embodiment of the present invention, in which an argon ion laser 21 has a wavelength of 363.8 nm, a diameter of ˜2 mm, and a single transverse mode (ie, having a Gaussian intensity profile). And emits a substantially monochromatic beam 22 in a multi-longitudinal mode. The light is plane polarized. After passing through the electronically operated shutter 23, the beam 22 is incident on a motorized variable attenuator 24, which is connected to a control system 46. The control system 46 continuously changes the intensity of the transmitted beam during the exposure operation or changes it step by step substantially continuously (ie, digitized to 16 levels, for example). Electric variable attenuators are commercially available from companies such as Metrolux Optische Messtechnik GmbH (particularly the product of Variable Dielectric Laser Beam Attenuators) and Del Mar Photonics Inc. (particularly the product of Diffractive Variable Attenuators). The transmitted beam from the variable attenuator 24 subsequently enters a quarter wave plate 26, which forms a circularly polarized beam. The diameter of this beam is magnified by a beam expander 28 with a set of lenses, so that the resulting beam intensity profile can be derived from a Gaussian distribution by using a beam transducer 30, from the center of the beam. Can be more easily converted to a substantially uniform distribution. Suitable beam transducers are commercially available, for example, from Moltech GmbH (particularly the product of piShaper). The output beam of the beam converter 30 passes through the second beam expander 32. This beam expander 32 forms a collimated beam, the uniform part of the center of which has a diameter larger than the size of the pattern to be printed. This beam is reflected by the mirror 34 onto the mask 38 so that the pattern 39 on the mask 38 is uniformly illuminated by the beam 35 collimated at normal incidence. On the lower surface of the mask, there is a two-dimensional periodic pattern of holes 39 in an opaque layer, which is arranged in a hexagonal lattice with a nearest neighbor distance of 520 nm. In the drawing, only five holes are shown in the mask pattern 13, but there are many orders of magnitude more holes, and the mask pattern 13 typically has dimensions of centimeters. It is understood. Pattern 39 is formed in a layer of chromium on a thick fused silica substrate using standard electron beam mask making techniques. The mask 38 is firmly fixed to a support frame (not shown). Below the mask 38 is a wafer 40 which is spin coated with a ˜1 μm thick layer of standard i-ray sensitive photoresist 41. The wafer 40 is mounted on the vacuum chuck 42. The vacuum chuck 42 is attached to a mechanical positioning system 44 that includes a plurality of actuators for positioning the wafer 40 substantially parallel to and in the vicinity of the pattern 39 in the mask 38. Collaborate with. The actuators preferably have three piezoelectric transducers (PZT), each of which has a closed-loop control of their respective displacements to minimize displacement errors caused by hysteresis and drift. It has an integrated strain gauge or capacitive sensor to be realized and advantageously has a long range of movement, for example 50 μm. The associated control system 46 for the actuator implements individual displacements or parallel displacements at a constant speed. The actuator can tilt the wafer 40 in an orthogonal plane. For example, the wafer 40 is adjusted parallel to the mask 38 and in the vicinity of the mask 38 using known equal thickness reference spacers inserted on different sides of the wafer 40. The initial separation between the wafer 40 and the mask 38 can generally be set to a value of 20 μm.

  Illumination of the hexagonal pattern of the structure 39 in the mask 38 using light of a wavelength of 363.8 nm generates a transmitted light field consisting of one zero order non-diffracted beam and six first order diffraction orders, This affects the formation of self-images separated by a Talbot distance of ˜0.88 μm. Since there are no second or higher order diffraction orders, the transmitted light field is exactly periodic in the direction orthogonal to the mask (ignoring the edges of the pattern). Since the illumination beam 35 is circularly polarized, the components of polarization in the orthogonal plane are equal, thereby realizing a symmetric distribution of diffraction orders and a symmetric structure in the self-image. Photoresist is applied from mask 38 using prior art DTL by wafer 40 being displaced longitudinally through the optical field by a distance corresponding to an integral multiple of the separation distance between successive Talbot surfaces. The time average intensity distribution recorded on the wafer 40 thus obtained can be obtained by computer simulation. The results are illustrated in FIG. 14, which shows a unit cell of a hexagonal array of intensity peaks having a nearest neighbor distance equal to the pattern 39 in the mask 38. Since the light field after passing through the mask 38 is exactly periodic in a direction perpendicular to the mask, the intensity of the peaks in this distribution does not depend on the average separation distance between the wafer 40 and the mask 38 during DTL exposure. . On the other hand, if the displacement distance is not exactly an integer multiple of the separation distance between the Talbot surfaces (eg due to mechanical hysteresis), the peak intensity of the distribution is no longer sensitive to the average separation distance. This dependence has been evaluated by computer simulation for displacement distances of 1.75 μm, 1.85 μm and 1.95 μm, respectively, and the results are shown in FIG. From this result, when the displacement is from twice the Talbot distance to ˜0.01 μm, the variation in peak intensity due to the change in the average separation distance is <1%, but from the double Talbot distance to 0.09 μm and For displacements of only 0.19 μm, the peak intensity varies ˜9% and ˜16%, respectively, due to variations in average separation, which is unacceptably large for some applications. The sensitivity of the printed pattern with respect to the absolute value of the displacement can be further evaluated by determining the dependence of the integrated intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the integrated distribution. FIG. 16 shows the result that it can be seen that the integrated intensity reaches the maximum value when the displacement is ˜1.25 μm, and continues to fluctuate greatly (only ± 14%) as the displacement distance further increases. ing.

  In this second embodiment of the invention, the wafer 40 is displaced longitudinally with respect to the mask 38 at a constant speed during exposure in accordance with prior art DTL practices. However, unlike the prior art, the intensity of the illumination beam 35 is not constant during exposure to record the average of the lateral intensity distribution between the Talbot surfaces, but instead, between the wafer 40 and the mask 38. The energy density of the irradiation at each increment of the separation distance is changed so as to change over the range of the separation distance. The intensity of the beam 35 while displacing the wafer 40 is controlled by the control system 46, which controls the transmission of the variable attenuator 24 to the energy density per separation displacement increment over a range of separation distances. Adjust according to the pre-programmed function that represents the desired variation of. The control system 46 advantageously opens and closes the shutter 23 at the beginning and end of exposure, respectively, to ensure that the photoresist 41 is not exposed. Advantageously, the aforementioned function substantially corresponds to a truncated Gaussian distribution as represented by equation (4). As in the first embodiment, the value σ assigned to the standard deviation of this function is half the separation of the Talbot plane T in the transmitted light field, and the maximum displacement of the wafer 40 during exposure is It is recommended that it be set to twice the Talbot distance (ie, t = 2), but other values can be used depending on the specific requirements of the relevant application. For σ = T / 2 and t = 2, the variation in irradiation dose for each displacement increment of the separation distance over the displacement distance is represented by the “Gaussian” curve in FIG. When the wafer 40 is exposed with a Gaussian distribution of energy density for each displacement increment of the separation distance between the mask 38 and the wafer 40, the sensitivity of the peak intensity in the integrated time average distribution to the fluctuation of the average separation distance during exposure and the desired The sensitivity to the deviation of the displacement distance of the wafer from this value can be evaluated by computer simulation. For σ = 0.4375 μm, 0.4625 μm and 0.4875 μm, and t = 2, the wafer displacement is 1.75 μm, 1.85 μm and 1.95 μm, respectively, and the results produced for the associated pattern are respectively It is shown in FIG. This allows the peak intensity variation due to the change in average separation to be estimated to be ˜2%, ˜0.6% and ˜0.9% for the three values of displacement, respectively. Is much lower than the corresponding value previously determined for DTL exposure. The sensitivity of the printed pattern with respect to the absolute value of the displacement can be further evaluated by determining the dependence of the integrated intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the integrated distribution. FIG. 18 shows that the average intensity reaches the maximum value when the displacement distance is about twice the Talbot distance (ie, 1.8 μm), and fluctuates with an amplitude of <± 1.5% as the distance further increases. The results you can see are shown. These residual variations can be further reduced by adjusting the parameters of the truncated Gaussian distribution, if necessary. The results show that exposure of the mask with exposure density per displacement increment that varies with the change in separation according to the truncated Gaussian distribution is much higher than can be achieved using prior art DTL. It has been demonstrated that printed patterns can be realized with uniformity and reproducibility.

  With the apparatus of this embodiment, another profile of the intensity change of the irradiating beam during exposure can alternatively be used, and similar beneficial results can be obtained. For example, a truncated sine profile can be used so that the resulting exposure energy density variation for each displacement increment of the wafer 40 over the displacement distance is represented by equation (6) for the first embodiment. It has such a truncated sine distribution. With respect to the embodiment, it is recommended that the displacement distance of the wafer 40 during exposure corresponds to twice the separation distance of the Talbot surface in the transmitted light field. For L = T and t = 1, the resulting variation in irradiation dose for each displacement increment of the separation distance over the displacement distance is represented by the “sine” curve in FIG. The computer simulation exposes the wafer 40, and the sensitivity of the peak intensity in the resulting integrated time average distribution to variations in the average separation distance between the wafer 40 and the mask 38, as well as the displacement distance of the wafer 40 from the desired value. The same is done to determine the sensitivity to the deviation. If L = 0.875 μm, 0.925 μm and 0.975 μm in equation (6) and t = 1, the displacement distance of the wafer is 1.75 μm, 1.85 μm and 1.95 μm, respectively, and the associated mask The resulting results for the patterns are shown in FIG. Thereby, the fluctuation of the peak intensity due to the change of the average separation distance can be estimated to be ˜0.3%, ˜2.5% and ˜4% for the three values of displacement, respectively. Is much lower than the corresponding value previously determined for DTL exposure. The sensitivity of the printed pattern with respect to the absolute value of the displacement can be further evaluated by determining the dependence of the integrated intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the integrated distribution. FIG. 20 shows that the average intensity reaches a maximum when the displacement is about twice the Talbot distance (ie 1.8 μm), and fluctuates with an amplitude of <± 1.5% as the distance further increases. The results can be seen. The results show that the exposure of the mask with an exposure density at each displacement increment that varies with the change in separation according to the sinusoidal distribution is much higher than what can be achieved using prior art DTL. It has been proved that a printed pattern can be realized with appearance and reproducibility.

  Another profile of intensity variation that can be used is triangular variation, whereby the resulting exposure energy density for each displacement increment of the wafer 40 over the displacement distance is given by equation (8) for the first embodiment. It has a triangular distribution as represented by For that embodiment, it is recommended that the maximum displacement of the wafer 40 during exposure correspond to twice the Talbot surface separation in the transmitted light field. For L = T and t = 1, the resulting variation of the radiation dose per displacement increment of the separation distance over the displacement distance is represented by the “triangular” curve in FIG. The computer simulation exposes the wafer 40, and the sensitivity of the peak intensity in the resulting integrated time average distribution to variations in the average separation distance between the wafer 40 and the mask 38, as well as the displacement distance of the wafer 40 from the desired value. A similar procedure is performed to assess sensitivity to deviation. When L = 0.875 μm, 0.925 μm and 0.975 μm in equation (8) and t = 1, the displacement distance of the wafer is 1.75 μm, 1.85 μm and 1.95 μm, respectively, and the associated mask The results produced for the pattern are shown in FIG. Thereby, the variation in peak intensity due to the change in average separation is estimated to be ~ 0.3%, ~ 0.6% and ~ 1.7% for the three values of displacement, respectively, and thus according to the prior art It is substantially lower than the corresponding value previously determined for DTL exposure. The sensitivity of the printed pattern with respect to the absolute value of the displacement can be further evaluated by determining the dependence of the integrated intensity on the displacement at a point having a lateral position corresponding to the intensity peak in the integrated distribution. FIG. 22 shows that the intensity reaches a maximum when the displacement is about twice the Talbot distance (ie 1.8 μm), and fluctuates with an amplitude of <± 1.5% as the distance further increases. The results that can be seen are shown. The results show that the exposure of the mask with an exposure density at each displacement increment that varies with the change in separation according to the triangular distribution results in a substantially higher uniformity than that obtained using prior art DTL. It was proved that a printed pattern can be realized with reproducibility.

  It can be seen from FIG. 6 that truncation Gaussian variation, truncation sinusoidal variation and truncation triangular variation of the incremental energy density are similar due to the change in separation distance described in this embodiment. That is, in each case, the full width of the profile is approximately twice the Talbot distance, and the full width at half-maximum (FWHM) is approximately the Talbot distance. In all cases, the incremental energy density changes substantially gradually over the displacement distance, and the minimum value of the initial and end values of the separation distance is> 80% of the maximum value. Thus, similar improvements in printed pattern uniformity and reproducibility can be achieved using properties similar to those shown, for example, an incremental dose variation profile with an appropriate trapezoidal distribution. Can be achieved. Furthermore, logical and experimental results show that higher uniformity and reproducibility of the printed pattern is the full width from the value suggested in FIG. 6 (with respect to the Talbot distance for the related periodic pattern). And by incrementing FWHM. However, larger displacements of the wafer 40 relative to the mask 38 during exposure require higher accuracy in the orthogonality of the wafer displacement relative to that surface (to avoid reducing print resolution) and have a greater range of movement ( One or more actuators are required and not necessarily desired.

  In general, to obtain very good uniformity and reproducibility of the printed results, the FWHM of the curve representing the change in exposure energy density for each displacement increment of the separation distance over the displacement distance is at least 0.8 of the Talbot distance. A value that is doubled but smaller, down to half the Talbot distance, can be used alternatively with lower uniformity and reproducibility results, and the distribution The total width should advantageously be at least 0.8 times twice the Talbot distance, but also smaller values that are reduced to the Talbot distance are used to accommodate relatively demanding applications be able to.

  With respect to the separation at which the maximum energy density per displacement increment occurs (as in the case of all curves shown in FIG. 6), the profile representing the variation in energy density per increment due to the change in separation is symmetrical. Or at least substantially symmetric. Thus, advantageously, the separation distance at which the maximum energy density per displacement increment occurs is at an intermediate point, or at least substantially at an intermediate point, between the initial value and the final value of the separation distance.

  Most advantageously, the initial value and final value of the separation distance between the mask 38 and the wafer 40 is at least 80% of the energy density value per displacement increment over the range of the separation distance. Few. However, a similar variation in energy density per displacement increment over the displacement distance can also be used, but still much better than the uniformity and reproducibility of patterns printed using prior art DTL exposure. Despite this, the uniformity and reproducibility of the printed result is low.

  In another embodiment of the present invention, the separation between the wafer and the mask is changed by twice the distance shown in FIG. 6 (ie, ˜4 times the Talbot distance), and every displacement increment during the displacement. Energy density is varied in response to the same truncated Gaussian / truncated sine / triangular profile shown in FIG. 6 (ie, with a full width of ~ 2 times the Talbot distance), but a relatively long displacement distance Continuously changed twice during the entire exposure. That is, although it is two exposures of the same printing pattern, it substantially corresponds to the two exposures of the printing pattern in which the separation distance is changed between the two exposures. Obviously, this exposure strategy can be extended to three or more sequential exposures.

  If the mask contains a quasi-periodic pattern, such as a chirped diffraction grating, that is, a pattern whose period changes "slowly" over the pattern area, the variation in exposure energy density for each displacement increment used to expose the pattern Should advantageously be based on the maximum period in the pattern.

  A two-dimensional periodic pattern (referred to as an “asymmetric” pattern) with components having different periods along different axes, eg, a plurality of holes arranged in a rectangular grid with different periods in the orthogonal direction For a given array of light, the light field transmitted through the mask generally does not form a regular self-image plane, and therefore generally a regular Talbot distance cannot be defined (Talbot for one pattern component). An exception is raised if the distance is a multiple of the Talbot distance for the other pattern component). In general, even when there is no self-image plane and the periodic pattern is not in the vertical direction, the application of the DTL technique can form a two-dimensional periodic pattern. However, due to non-periodic variations in the diffraction pattern, it is generally difficult to achieve a steady distribution without using a DTL range. This DTL range is much larger than needed for that period if the required period exists separately. For those cases, the profile of the energy density variation used while the separation is changing is obtained by computer simulation of the integrated intensity distribution formed on the wafer under different exposure conditions for the particular mask pattern involved. It is recommended to seek. In general, for periodic, quasi-periodic or asymmetric patterns, the exposure energy density per displacement increment varies substantially stepwise over the displacement distance, preferably at least 50%. By using the pattern printed on the recording layer, the distance from the desired value and the initial value of the separation distance compared to the pattern printed using the DTL according to the prior art. Much lower sensitivity.

  In the first embodiment described above, a one-dimensional periodic pattern is exposed using a variable speed of displacement of the wafer 40, whereas in the second embodiment, irradiation is performed. It is understood that the two-dimensional pattern is exposed using a variable intensity, but the exposure strategy used is of course independent of the type of array. That is, the variable velocity scheme of displacement can be applied equally well to a two-dimensional pattern, and the variable intensity scheme can be applied equally well to a one-dimensional pattern.

  In the second embodiment, the displacement of the wafer 40 relative to the mask 38 is achieved by displacing the wafer 40, but in another embodiment of the present invention, alternatively, the mask 38 during exposure. It is understood that the same effect and printing result can be achieved by displacing the image in the vertical direction. To this end, a suitable mechanical system that cooperates with one or more actuators and associated control systems for displacing the mask 38 should be incorporated into the apparatus.

  In another embodiment of the invention, the same (or equivalent) device is used in the second embodiment, but the variable attenuator 24 is not used. Since the wafer 40 is displaced relative to the mask 38 at a constant speed, the illumination system rather exposes the mask 38 to the beam in a series of “sub-exposures” having sub-exposure frequencies. In this sub-exposure, the instantaneous intensity of the beam is equal for each sub-exposure, but the exposure time per sub-exposure is measured over a unit time that includes the effective intensity (ie, at least one period of multiple pulses). Is continuously changed (ie, the duty cycle of the pulsed exposure is changed). Such effective intensity varies with changes in the separation distance to produce the required variation in energy density for each displacement increment of the separation distance over the displacement distance. The frequency of the sub-exposure should be high enough with respect to the speed of mask displacement to produce the desired effect. By doing so, the time-averaged intensity distribution irradiating the photoresist-coated wafer (over full exposure) approximates the intensity distribution generated in the second embodiment, and so on. The printed results are substantially equal. The start time and end time of each sub-exposure can be defined by the shutter 23 and the control system 46 of the second embodiment.

  In a related embodiment, instead, the part of the shutter that blocks the beam is selected to be a partially transmissive material so that the intensity of the beam does not drop to zero when the shutter is closed. However, it drops to a relatively low value and has a relatively high value when the shutter is opened. By selecting different exposure parameters such that the required variation in energy density for each displacement increment of separation is substantially obtained over the displacement distance, a substantially identical time-average intensity distribution can be obtained in the photoresist. (It is clear that the above-described embodiment can be obtained if the shutter is completely opaque and a relatively low value can be considered as 0).

  In another related embodiment, the same exposure time is used for each sub-exposure, and instead the beam intensity is varied over a series of sub-exposures, thereby reducing the separation distance over the displacement distance. By producing the required fluctuations in energy density for each displacement increment, substantially the same printing results can be achieved.

  These variations occur when the light intensity is defined in terms of exposure energy per i unit time (however, the time unit is defined as the period of sub-exposure). The second embodiment is the same as the second embodiment in that the same fluctuation in intensity is caused by the change in the separation distance that may be caused.

  In another related embodiment, a pulsed laser is used as the illumination source to obtain the required variation in energy density for each displacement increment of the separation distance over the displacement distance, and the unit time (corresponding to multiple pulses) The intensity of the beam (measured over time) is changed by changing the frequency of the pulse (and by using a substantially constant energy per pulse). The laser light source used in the first and second embodiments is an argon laser operating at a specific wavelength of 363.8 nm and emitting a continuous wave (CW) beam, although another implementation of the present invention. In this embodiment, alternative laser light sources that emit beams at different wavelengths, such as solid state lasers, laser diodes and excimer lasers, can be used. Advantageously, the output beam has a UV wavelength, so that a conventional UV-sensitive photoresist can be used, although a beam having a wavelength in another part of the electromagnetic spectrum is also an option. Can be used. The laser can also supply light in pulses, and advantageously can provide light having a higher frequency (eg, ≧ kHz) rather than CW, as is common with solid state diode pumped lasers. . In another embodiment, another type of light source for generating a substantially monochromatic light beam, for example, a filter for separating the required spectral components and a well collimated beam is formed. Arc lamps (eg mercury lamps) in combination with suitable optical elements can be used.

  While the first and second embodiments use a refractive beam converter to form a beam having a substantially uniform intensity from an output laser beam having a Gaussian profile, In embodiments, other means can be used to achieve substantially uniform illumination of the mask. For example, if the laser beam has a Gaussian intensity profile, the laser beam can be expanded and subsequently scanned with a raster pattern across the mask so that the time averaged energy density is consistent across the pattern. It is made like. When using such a scan strategy, it is necessary to repeatedly change the mask-wafer separation distance over the displacement distance while the energy density per displacement increment changes between each change, and the separation distance. The frequency of the change needs to be high with respect to the frequency of the scan lines in the raster pattern so that each part of the mask pattern is exposed in substantially the same way.

  In the second embodiment, the fluctuation of the intensity of the irradiation beam is caused by a variable attenuator introduced in the beam path following the laser. In another embodiment of the present invention, the fluctuation of the intensity of the beam is reduced. This can be achieved by other means, for example by modulating the drive current of the laser light source so that the power of the output beam from the laser light source is changed.

  In another embodiment of the invention, an immersion liquid, such as water, is used between the substrate and the mask to reduce the minimum period of the pattern that can be printed using a technique that uses a specific illumination wavelength. Can be included in the gap.

  The embodiments described above are considered to be advantageous embodiments of the present invention, but of course, it is understood that various modifications and changes in shape or details can be easily made without departing from the scope of the present invention. Is done. Accordingly, the present invention is not limited to the precise shape described above and illustrated, but is intended to cover any modifications falling within the scope of the appended claims.

Claims (20)

In a method of printing a pattern of structure,
a) preparing a substrate provided with a recording layer on the surface;
b) providing a mask comprising at least one of a periodic pattern of structures and a quasi-periodic pattern of structures;
c) placing the substrate substantially parallel to the mask and with an initial separation having a first desired value;
d) providing an illumination system for irradiating the mask with substantially monochromatic, substantially collimated light, having an intensity such that a light field transmitted through the mask exposes the recording layer; Steps,
e) illuminating the mask for an exposure time while changing the separation distance by a second desired value and the separation rate change rate, the separation rate change rate and light Irradiating the mask with an energy density for each incremental change in the separation distance that changes at least one of the intensities during the change of the separation distance, thereby changing with the change in the separation distance; Has
The pattern printed on the recording layer is for a deviation of the initial separation from the first desired value and for a deviation of the change in the separation from the second desired value. And a low sensitivity.   The method of claim 1, wherein the energy density for each incremental change in the separation is changed by at least 50% over the change in the separation. The transmitted light field forms a self-image plane separated by a Talbot distance;
The method of claim 1, wherein the change in the separation distance is greater than the Talbot distance. The transmitted light field forms a self-image plane separated by a Talbot distance;
The method of claim 1, wherein a variation in energy density for each incremental change in the separation distance across the separation distance has a full width at half maximum greater than half of the Talbot distance.   The energy density variation for each incremental change in the separation distance over the change in the separation distance substantially corresponds to one of a truncated Gaussian distribution, a truncated sine distribution, and a triangular distribution. Method. Continuously changing the separation distance;
The method of claim 1, wherein the rate of change in the separation distance represents a rate of displacement of the mask or the substrate. Changing the separation distance in successive steps having a step distance and a step frequency;
The method of claim 1, wherein if the rate of change of the separation distance is changed, at least one of the step distance and the step frequency is changed over the change of the separation distance.   The method of claim 1, wherein the light intensity is continuously changed over the change in the separation distance. The rate of change in the separation is substantially constant;
Changing the instantaneous intensity of the light multiple times between an upper limit value and a lower limit value for each incremental change in the separation distance;
The method of claim 1, wherein the intensity during each incremental change in the separation distance corresponds to a time average value. Changing the separation distance multiple times during the irradiation over the change in the separation distance;
The method of claim 1, wherein at least one of the rate of change in the separation distance and the intensity of light is changed during each change in the separation distance. In an apparatus for printing a structure pattern,
a) a substrate provided with a recording layer on its surface;
b) a mask comprising at least one of a periodic pattern of structures and a quasi-periodic pattern of structures;
c) means for positioning the substrate substantially parallel to the mask and at an initial separation having a first desired value;
d) an illumination system for illuminating the mask with a substantially monochromatic, substantially collimated light for an exposure time period, having an intensity such that a light field transmitted through the mask exposes the recording layer When,
e) means for changing the separation distance by a second desired value and at a rate of change of the separation distance;
f) during the irradiation of the mask such that the mask is irradiated with an energy density for each incremental change in the separation distance, which is changed gradually and gradually with the change in the separation distance; Means for changing at least one of a rate of change of distance and said intensity of light,
The pattern printed on the recording layer is for a deviation of the initial separation from the first desired value and for a deviation of the change in the separation from the second desired value. And a low sensitivity.   The changing means includes an actuator that changes the separation distance at a variable speed or changes the separation distance in successive steps with a period between steps changed with the change in the separation distance. Item 12. The apparatus according to Item 11.   12. The apparatus of claim 11, wherein the illumination system includes a light source that emits an output beam having an output, and the modifying means changes the output of the output beam.   The apparatus of claim 11, wherein the illumination system includes a variable attenuator that varies the intensity of light that illuminates the mask.   At least one of the periodic pattern and quasi-periodic pattern of the structure in the mask is formed in at least one of a layer of opaque material and a layer of phase shift material on a transparent substrate; The apparatus of claim 11.   The apparatus of claim 11, wherein at least one of the periodic pattern of the structure and the quasi-periodic pattern of the structure is periodic in at least one direction. The mask comprises a plurality of periodic sub-patterns;
12. At least one of the plurality of periodic subpatterns has a first period, and at least one of the plurality of periodic subpatterns has a different second period. The device described. The periodic pattern in the mask has a lattice vector oriented in a first direction;
The apparatus of claim 11, wherein the mask comprises at least one additional periodic pattern having a lattice vector oriented in a different second direction.   The apparatus of claim 11, wherein the illumination system includes a scanning system for forming a beam of light and scanning the beam across the mask.   The method of claim 1, further comprising using computer simulation to determine a required variation in energy density for each incremental change in the separation over the change in separation.

Images